Twelve years after a motorcycle crash left him paralyzed, Gert-Jan Oskam stood up, walked across a room, and climbed a set of stairs. He did it by thinking about it. No joystick, no button, no robot suit doing the work for him. Just his own intention to move, picked up by implants in his brain, decoded by artificial intelligence, and delivered straight to his spinal cord in a fraction of a second.
Honestly, the first time I watched the footage, I had to replay it. A man who was told he would never walk again, taking natural steps because a computer learned to read what he wanted his legs to do. Researchers call it a “digital bridge,” and it may be one of the most important things AI has ever done for the human body.
Digital Bridge: The Breakthrough in One Sentence
The digital bridge is a wireless brain spine interface that skips over the damaged part of the spinal cord entirely. Implants on the surface of the brain record the electrical activity that happens when Oskam thinks about moving. An AI decoder translates those signals into specific movement intentions in real time. A second implant on the spinal cord then stimulates the exact nerves that control his hips, knees, and ankles. The injury is still there. The message simply takes a different road.
The work was led by neuroscientist Grégoire Courtine of the Swiss Federal Institute of Technology in Lausanne (EPFL) and neurosurgeon Jocelyne Bloch of Lausanne University Hospital, together with engineers at Clinatec, part of the French research agency CEA, which developed the brain implants. The results were published in the journal Nature in May 2023.
Key Facts at a Glance
- Who: Gert-Jan Oskam, a Dutch man paralyzed in a 2011 motorcycle accident
- What: A wireless “digital bridge” between his brain and spinal cord
- Where the AI fits: Adaptive AI algorithms decode his movement intentions from brain signals in real time
- What he can do: Stand, walk, climb stairs, and move over uneven ground
- Surprise result: He regained some movement even with the system switched off
- Since then: The approach is being tested in more patients and adapted to restore arm and hand movement
Why a Spinal Cord Injury Stops Movement
To understand why this matters, it helps to picture how walking normally works. When you decide to take a step, neurons in the motor cortex, a strip of brain tissue running across the top of your head, fire off a command. That command travels down the spinal cord like a message down a cable. Lower in the spine, it reaches networks of neurons that coordinate the muscles in your legs.
A spinal cord injury cuts or damages that cable. Here’s the key thing many people don’t realize: the brain usually still works fine, and so do the nerve circuits below the injury. The problem is the broken connection in the middle. The brain keeps sending the message, but it never arrives.
That’s exactly the gap the digital bridge was designed to close.
How the Digital Bridge Works, Step by Step
Step 1: Brain Implants Listen to Movement Thoughts
Surgeons placed two disc shaped implants into openings in Oskam’s skull, resting over the areas of his brain that control leg movement. These devices use a technology called WIMAGINE, developed at Clinatec in Grenoble, France. Each one carries a grid of electrodes that sits on the surface of the brain rather than poking into it.
The implants record a type of signal called electrocorticography, or ECoG. Think of it as a high resolution recording of the brain’s electrical chatter, captured right where movement is planned. Because the electrodes don’t penetrate brain tissue, this approach is considered more stable over long periods than some other brain implants.
Step 2: Signals Travel Wirelessly to a Portable Computer
The brain implants don’t have wires running out of the head. Instead, they send their recordings wirelessly to antennas built into a custom headset. From there, the data goes to a portable processing unit that Oskam carries in a backpack. That small computer is where the real magic happens.
Step 3: AI Decodes What He Wants to Do
This is the heart of the system, and it’s where artificial intelligence earns its place in the story.
Raw brain signals are messy. They’re full of noise, they change from day to day, and every person’s brain is wired a little differently. You can’t just look at a spike of activity and say, “That means lift the left knee.” So the team used adaptive AI algorithms to learn the pattern.
During calibration, researchers asked Oskam to try to move specific joints, like flexing his hip or lifting his ankle. While he attempted each movement, the system recorded his brain activity. The AI then learned which patterns of brain signals matched which intended movements. Over time, it built a personal map of his movement thoughts.
Once trained, the decoder works continuously, predicting what he wants to do many times every second. It figures out not only which joint he wants to move, but also how strongly, which is why his steps look natural instead of robotic.
The word “adaptive” matters here. Brain signals drift, so the algorithm keeps updating itself instead of relying on a fixed rulebook. According to the researchers, the system can be calibrated within just a few minutes. Oskam himself said that within five to ten minutes, he could control his hips.
Step 4: Spinal Stimulation Turns Intention Into Movement
The decoded predictions are sent to a second implant, a neurostimulator placed over the region of the spinal cord that controls the legs. This device sends carefully timed pulses of electricity through electrodes resting on the spinal cord, a technique called epidural electrical stimulation.
Those pulses activate the groups of neurons below the injury that drive specific leg muscles. Because the stimulation is triggered by Oskam’s own decoded thoughts, the timing and strength match what he actually wants. The result is movement that’s voluntary, smooth, and under his control.
The whole loop, from thought to brain signal to AI decoding to spinal stimulation, happens fast enough to feel natural.
What Makes This Different From Earlier Treatments
Oskam wasn’t new to spinal stimulation. Years earlier, he joined a previous study by Courtine’s team that used stimulation alone, following preprogrammed patterns. It helped, but after three years his progress plateaued. The movements were triggered by the system, not by him, so they felt mechanical.
The digital bridge changed that. By adding the brain implant and the AI decoder, the stimulation now follows his intentions. He described being able to control his hips and feel like he was the one moving, not the machine. That sense of control is what researchers mean when they call this movement “natural.”
The Surprise: Healing Even When It’s Turned Off
Here’s the part that made a lot of scientists sit up. After months of training with the digital bridge, Oskam showed improvement even when the system was switched off. With crutches, he could walk on his own.
The researchers believe repeated use of the bridge helped his nervous system build new connections, a process called neurological recovery. In other words, the technology may not only replace the broken link but also help the body partially repair it. That’s a big deal, and it’s something the team is now studying closely.
Real Life, Not Just a Lab Demo
One detail I love: this wasn’t limited to a treadmill in a hospital. Oskam used the system in daily life. He could stand at a bar with friends and enjoy a beer, something he called a simple pleasure he had missed for years. Small moments like that are what this kind of research is really about.
What’s Happened Since
The original study involved just one person, and the researchers were clear about that. Since then, the work has kept moving.
The team has developed a version of the digital bridge for the cervical spinal cord, the part of the spine in the neck, with the goal of restoring arm and hand movement for people with tetraplegia, meaning paralysis from the neck down. Being able to grab a cup or feed yourself could be life changing for millions of people.
An EU funded project is also working toward a fully implantable brain spine interface, bringing together experts in neuroscience, bioengineering, and AI from the Netherlands, Switzerland, and France. The goal is a system that’s smaller, more practical, and eventually something patients could live with every day, without a backpack.
The researchers have also partnered with the medical technology company ONWARD Medical to shrink the system into a smaller device for clinical use. In 2024, ONWARD signed an agreement with the French research agency CEA for exclusive rights to the WIMAGINE brain implant technology, a sign that the digital bridge is moving from the lab toward real treatment.
The Challenges Ahead
As exciting as this is, it’s not a cure you can get at your local hospital yet. The system requires brain surgery and spinal surgery, which carry real risks. It’s expensive, it needs specialized teams to set up and calibrate, and it has only been tested in a small number of people. Large clinical trials will be needed before regulators can approve it for wider use. That process usually takes years.
There are also big questions about long term reliability, battery life, and how well the approach works for different types of spinal cord injuries.
Why This Matters for the Future of AI in Medicine
When people talk about AI, they often think of chatbots or image generators. The digital bridge shows a very different side of the technology. Here, AI acts as a translator between the brain and the body, learning one person’s unique signals and turning them into movement.
For anyone living with paralysis, and for the families who love them, that’s not science fiction anymore. It’s a real man walking up real stairs, powered by his own thoughts. And if the next stage succeeds, the same idea could help people hold a loved one’s hand again.
That’s the kind of AI story worth paying attention to.
Leave a Reply